Muutke küpsiste eelistusi

Machine Learning with Python for Everyone [Pehme köide]

4.43/5 (11 hinnangut Goodreads-ist)
  • Formaat: Paperback / softback, 592 pages, kõrgus x laius x paksus: 229x178x28 mm, kaal: 930 g
  • Sari: Addison-Wesley Data & Analytics Series
  • Ilmumisaeg: 17-Dec-2019
  • Kirjastus: Addison Wesley
  • ISBN-10: 0134845625
  • ISBN-13: 9780134845623
Teised raamatud teemal:
Machine Learning with Python for EveryoneSuurem pilt
  • Formaat: Paperback / softback, 592 pages, kõrgus x laius x paksus: 229x178x28 mm, kaal: 930 g
  • Sari: Addison-Wesley Data & Analytics Series
  • Ilmumisaeg: 17-Dec-2019
  • Kirjastus: Addison Wesley
  • ISBN-10: 0134845625
  • ISBN-13: 9780134845623
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9780134845623.html
Märksõnad:
The Complete Beginner's Guide to Understanding and Building Machine Learning Systems with Python

Machine Learning with Python for Everyone will help you master the processes, patterns, and strategies you need to build effective learning systems, even if you're an absolute beginner. If you can write some Python code, this book is for you, no matter how little college-level math you know. Principal instructor Mark E. Fenner relies on plain-English stories, pictures, and Python examples to communicate the ideas of machine learning.

Mark begins by discussing machine learning and what it can do; introducing key mathematical and computational topics in an approachable manner; and walking you through the first steps in building, training, and evaluating learning systems. Step by step, you'll fill out the components of a practical learning system, broaden your toolbox, and explore some of the field's most sophisticated and exciting techniques. Whether you're a student, analyst, scientist, or hobbyist, this guide's insights will be applicable to every learning system you ever build or use.



Understand machine learning algorithms, models, and core machine learning concepts Classify examples with classifiers, and quantify examples with regressors Realistically assess performance of machine learning systems Use feature engineering to smooth rough data into useful forms Chain multiple components into one system and tune its performance Apply machine learning techniques to images and text Connect the core concepts to neural networks and graphical models Leverage the Python scikit-learn library and other powerful tools

Register your book for convenient access to downloads, updates, and/or corrections as they become available. See inside book for details.
Foreword xxi
Preface xxiii
About the Author xxvii
I First Steps 1(106)
1 Let's Discuss Learning
3(16)
1.1 Welcome
3(1)
1.2 Scope, Terminology, Prediction, and Data
4(3)
1.2.1 Features
5(1)
1.2.2 Target Values and Predictions
6(1)
1.3 Putting the Machine in Machine Learning
7(2)
1.4 Examples of Learning Systems
9(2)
1.4.1 Predicting Categories: Examples of Classifiers
9(1)
1.4.2 Predicting Values: Examples of Regressors
10(1)
1.5 Evaluating Learning Systems
11(2)
1.5.1 Correctness
11(1)
1.5.2 Resource Consumption
12(1)
1.6 A Process for Building Learning Systems
13(2)
1.7 Assumptions and Reality of Learning
15(2)
1.8 End-of-Chapter Material
17(2)
1.8.1 The Road Ahead
17(1)
1.8.2 Notes
17(2)
2 Some Technical Background
19(36)
2.1 About Our Setup
19(1)
2.2 The Need for Mathematical Language
19(1)
2.3 Our Software for Tackling Machine Learning
20(1)
2.4 Probability
21(7)
2.4.1 Primitive Events
22(1)
2.4.2 Independence
23(1)
2.4.3 Conditional Probability
24(1)
2.4.4 Distributions
25(3)
2.5 Linear Combinations, Weighted Sums, and Dot Products
28(6)
2.5.1 Weighted Average
30(2)
2.5.2 Sums of Squares
32(1)
2.5.3 Sum of Squared Errors
33(1)
2.6 A Geometric View: Points in Space
34(9)
2.6.1 Lines
34(5)
2.6.2 Beyond Lines
39(4)
2.7 Notation and the Plus-One Trick
43(2)
2.8 Getting Groovy, Breaking the Straight-Jacket, and Nonlinearity
45(2)
2.9 NumPy versus "All the Maths"
47(5)
2.9.1 Back to 1D versus 2D
49(3)
2.10 Floating-Point Issues
52(1)
2.11 EOC
53(2)
2.11.1 Summary
53(1)
2.11.2 Notes
54(1)
3 Predicting Categories: Getting Started with Classification
55(30)
3.1 Classification Tasks
55(1)
3.2 A Simple Classification Dataset
56(3)
3.3 Training and Testing: Don't Teach to the Test
59(3)
3.4 Evaluation: Grading the Exam
62(1)
3.5 Simple Classifier #1: Nearest Neighbors, Long Distance Relationships, and Assumptions
63(5)
3.5.1 Defining Similarity
63(1)
3.5.2 The k in k-NN
64(1)
3.5.3 Answer Combination
64(1)
3.5.4 k-NN, Parameters, and Nonparametric Methods
65(1)
3.5.5 Building a k-NN Classification Model
66(2)
3.6 Simple Classifier #2: Naive Bayes, Probability, and Broken Promises
68(2)
3.7 Simplistic Evaluation of Classifiers
70(11)
3.7.1 Learning Performance
70(1)
3.7.2 Resource Utilization in Classification
71(6)
3.7.3 Stand-Alone Resource Evaluation
77(4)
3.8 EOC
81(4)
3.8.1 Sophomore Warning: Limitations and Open Issues
81(1)
3.8.2 Summary
82(1)
3.8.3 Notes
82(1)
3.8.4 Exercises
83(2)
4 Predicting Numerical Values: Getting Started with Regression
85(22)
4.1 A Simple Regression Dataset
85(2)
4.2 Nearest-Neighbors Regression and Summary Statistics
87(4)
4.2.1 Measures of Center: Median and Mean
88(2)
4.2.2 Building a k-NN Regression Model
90(1)
4.3 Linear Regression and Errors
91(7)
4.3.1 No Flat Earth: Why We Need Slope
92(2)
4.3.2 Tilting the Field
94(3)
4.3.3 Performing Linear Regression
97(1)
4.4 Optimization: Picking the Best Answer
98(3)
4.4.1 Random Guess
98(1)
4.4.2 Random Step
99(1)
4.4.3 Smart Step
99(1)
4.4.4 Calculated Shortcuts
100(1)
4.4.5 Application to Linear Regression
101(1)
4.5 Simple Evaluation and Comparison of Regressors
101(3)
4.5.1 Root Mean Squared Error
101(1)
4.5.2 Learning Performance
102(1)
4.5.3 Resource Utilization in Regression
102(2)
4.6 EOC
104(5)
4.6.1 Limitations and Open Issues
104(1)
4.6.2 Summary
105(1)
4.6.3 Notes
105(1)
4.6.4 Exercises
105(2)
II Evaluation 107(128)
5 Evaluating and Comparing Learners
109(50)
5.1 Evaluation and Why Less Is More
109(1)
5.2 Terminology for Learning Phases
110(6)
5.2.1 Back to the Machines
110(3)
5.2.2 More Technically Speaking
113(3)
5.3 Major Tom, There's Something Wrong: Overfitting and Underfitting
116(9)
5.3.1 Synthetic Data and Linear Regression
117(1)
5.3.2 Manually Manipulating Model Complexity
118(2)
5.3.3 Goldilocks: Visualizing Overfitting, Underfitting, and "Just Right"
120(4)
5.3.4 Simplicity
124(1)
5.3.5 Take-Home Notes on Overfitting
124(1)
5.4 From Errors to Costs
125(3)
5.4.1 Loss
125(1)
5.4.2 Cost
126(1)
5.4.3 Score
127(1)
5.5 (Re)Sampling: Making More from Less
128(14)
5.5.1 Cross-Validation
128(4)
5.5.2 Stratification
132(1)
5.5.3 Repeated Train-Test Splits
133(4)
5.5.4 A Better Way and Shuffling
137(3)
5.5.5 Leave-One-Out Cross-Validation
140(2)
5.6 Break-It-Down: Deconstructing Error into Bias and Variance
142(7)
5.6.1 Variance of the Data
143(1)
5.6.2 Variance of the Model
144(1)
5.6.3 Bias of the Model
144(1)
5.6.4 All Together Now
145(1)
5.6.5 Examples of Bias-Variance Tradeoffs
145(4)
5.7 Graphical Evaluation and Comparison
149(5)
5.7.1 Learning Curves: How Much Data Do We Need?
150(2)
5.7.2 Complexity Curves
152(2)
5.8 Comparing Learners with Cross-Validation
154(1)
5.9 EOC
155(4)
5.9.1 Summary
155(1)
5.9.2 Notes
155(2)
5.9.3 Exercises
157(2)
6 Evaluating Classifiers
159(46)
6.1 Baseline Classifiers
159(2)
6.2 Beyond Accuracy: Metrics for Classification
161(9)
6.2.1 Eliminating Confusion from the Confusion Matrix
163(1)
6.2.2 Ways of Being Wrong
164(1)
6.2.3 Metrics from the Confusion Matrix
165(1)
6.2.4 Coding the Confusion Matrix
166(2)
6.2.5 Dealing with Multiple Classes: Multiclass Averaging
168(2)
6.2.6 F1
170(1)
6.3 ROC Curves
170(11)
6.3.1 Patterns in the ROC
173(1)
6.3.2 Binary ROC
174(3)
6.3.3 AUC: Area-Under-the-(ROC)-Curve
177(2)
6.3.4 Multiclass Learners, One-versus-Rest, and ROC
179(2)
6.4 Another Take on Multiclass: One-versus-One
181(4)
6.4.1 Multiclass AUC Part Two: The Quest for a Single Value
182(3)
6.5 Precision-Recall Curves
185(2)
6.5.1 A Note on Precision-Recall Tradeoff
185(1)
6.5.2 Constructing a Precision-Recall Curve
186(1)
6.6 Cumulative Response and Lift Curves
187(3)
6.7 More Sophisticated Evaluation of Classifiers: Take Two
190(11)
6.7.1 Binary
190(5)
6.7.2 A Novel Multiclass Problem
195(6)
6.8 EOC
201(4)
6.8.1 Summary
201(1)
6.8.2 Notes
202(1)
6.8.3 Exercises
203(2)
7 Evaluating Regressors
205(30)
7.1 Baseline Regressors
205(2)
7.2 Additional Measures for Regression
207(7)
7.2.1 Creating Our Own Evaluation Metric
207(1)
7.2.2 Other Built-in Regression Metrics
208(1)
7.2.3 R2
209(5)
7.3 Residual Plots
214(7)
7.3.1 Error Plots
215(2)
7.3.2 Residual Plots
217(4)
7.4 A First Look at Standardization
221(4)
7.5 Evaluating Regressors in a More Sophisticated Way: Take Two
225(7)
7.5.1 Cross-Validated Results on Multiple Metrics
226(4)
7.5.2 Summarizing Cross-Validated Results
230(1)
7.5.3 Residuals
230(2)
7.6 EOC
232(5)
7.6.1 Summary
232(1)
7.6.2 Notes
232(2)
7.6.3 Exercises
234(1)
III More Methods and Fundamentals 235(150)
8 More Classification Methods
237(58)
8.1 Revisiting Classification
237(2)
8.2 Decision Trees
239(10)
8.2.1 Tree-Building Algorithms
242(3)
8.2.2 Let's Go: Decision Tree Time
245(4)
8.2.3 Bias and Variance in Decision Trees
249(1)
8.3 Support Vector Classifiers
249(10)
8.3.1 Performing SVC
253(3)
8.3.2 Bias and Variance in SVCs
256(3)
8.4 Logistic Regression
259(10)
8.4.1 Betting Odds
259(3)
8.4.2 Probabilities, Odds, and Log-Odds
262(5)
8.4.3 Just Do It: Logistic Regression Edition
267(1)
8.4.4 A Logistic Regression: A Space Oddity
268(1)
8.5 Discriminant Analysis
269(16)
8.5.1 Covariance
270(12)
8.5.2 The Methods
282(1)
8.5.3 Performing DA
283(2)
8.6 Assumptions, Biases, and Classifiers
285(2)
8.7 Comparison of Classifiers: Take Three
287(3)
8.7.1 Digits
287(3)
8.8 EOC
290(5)
8.8.1 Summary
290(1)
8.8.2 Notes
290(3)
8.8.3 Exercises
293(2)
9 More Regression Methods
295(26)
9.1 Linear Regression in the Penalty Box: Regularization
295(6)
9.1.1 Performing Regularized Regression
300(1)
9.2 Support Vector Regression
301(7)
9.2.1 Hinge Loss
301(4)
9.2.2 From Linear Regression to Regularized Regression to Support Vector Regression
305(2)
9.2.3 Just Do It-SVR Style
307(1)
9.3 Piecewise Constant Regression
308(5)
9.3.1 Implementing a Piecewise Constant Regressor
310(1)
9.3.2 General Notes on Implementing Models
311(2)
9.4 Regression Trees
313(1)
9.4.1 Performing Regression with Trees
313(1)
9.5 Comparison of Regressors: Take Three
314(4)
9.6 EOC
318(3)
9.6.1 Summary
318(1)
9.6.2 Notes
318(1)
9.6.3 Exercises
319(2)
10 Manual Feature Engineering: Manipulating Data for Fun and Profit
321(38)
10.1 Feature Engineering Terminology and Motivation
321(3)
10.1.1 Why Engineer Features?
322(1)
10.1.2 When Does Engineering Happen?
323(1)
10.1.3 How Does Feature Engineering Occur?
324(1)
10.2 Feature Selection and Data Reduction: Taking out the Trash
324(1)
10.3 Feature Scaling
325(4)
10.4 Discretization
329(3)
10.5 Categorical Coding
332(9)
10.5.1 Another Way to Code and the Curious Case of the Missing Intercept
334(7)
10.6 Relationships and Interactions
341(9)
10.6.1 Manual Feature Construction
341(2)
10.6.2 Interactions
343(5)
10.6.3 Adding Features with Transformers
348(2)
10.7 Target Manipulations
350(6)
10.7.1 Manipulating the Input Space
351(2)
10.7.2 Manipulating the Target
353(3)
10.8 EOC
356(3)
10.8.1 Summary
356(1)
10.8.2 Notes
356(1)
10.8.3 Exercises
357(2)
11 Tuning Hyperparameters and Pipelines
359(26)
11.1 Models, Parameters, Hyperparameters
360(2)
11.2 Tuning Hyperparameters
362(8)
11.2.1 A Note on Computer Science and Learning Terminology
362(1)
11.2.2 An Example of Complete Search
362(6)
11.2.3 Using Randomness to Search for a Needle in a Haystack
368(2)
11.3 Down the Recursive Rabbit Hole: Nested Cross-Validation
370(7)
11.3.1 Cross-Validation, Redux
370(1)
11.3.2 GridSearch as a Model
371(1)
11.3.3 Cross-Validation Nested within Cross-Validation
372(3)
11.3.4 Comments on Nested CV
375(2)
11.4 Pipelines
377(3)
11.4.1 A Simple Pipeline
378(1)
11.4.2 A More Complex Pipeline
379(1)
11.5 Pipelines and Tuning Together
380(2)
11.6 EOC
382(5)
11.6.1 Summary
382(1)
11.6.2 Notes
382(1)
11.6.3 Exercises
383(2)
IV Adding Complexity 385(144)
12 Combining Learners
387(22)
12.1 Ensembles
387(2)
12.2 Voting Ensembles
389(1)
12.3 Bagging and Random Forests
390(8)
12.3.1 Bootstrapping
390(4)
12.3.2 From Bootstrapping to Bagging
394(2)
12.3.3 Through the Random Forest
396(2)
12.4 Boosting
398(3)
12.4.1 Boosting Details
399(2)
12.5 Comparing the Tree-Ensemble Methods
401(4)
12.6 EOC
405(4)
12.6.1 Summary
405(1)
12.6.2 Notes
405(1)
12.6.3 Exercises
406(3)
13 Models That Engineer Features for Us
409(60)
13.1 Feature Selection
411(17)
13.1.1 Single-Step Filtering with Metric-Based Feature Selection
412(11)
13.1.2 Model-Based Feature Selection
423(3)
13.1.3 Integrating Feature Selection with a Learning Pipeline
426(2)
13.2 Feature Construction with Kernels
428(17)
13.2.1 A Kernel Motivator
428(5)
13.2.2 Manual Kernel Methods
433(5)
13.2.3 Kernel Methods and Kernel Options
438(4)
13.2.4 Kernelized SVCs: SVMs
442(1)
13.2.5 Take-Home Notes on SVM and an Example
443(2)
13.3 Principal Components Analysis: An Unsupervised Technique
445(17)
13.3.1 A Warm Up: Centering
445(3)
13.3.2 Finding a Different Best Line
448(1)
13.3.3 A First PCA
449(3)
13.3.4 Under the Hood of PCA
452(5)
13.3.5 A Finale: Comments on General PCA
457(1)
13.3.6 Kernel PCA and Manifold Methods
458(4)
13.4 EOC
462(7)
13.4.1 Summary
462(1)
13.4.2 Notes
462(5)
13.4.3 Exercises
467(2)
14 Feature Engineering for Domains: Domain-Specific Learning
469(28)
14.1 Working with Text
470(9)
14.1.1 Encoding Text
471(5)
14.1.2 Example of Text Learning
476(3)
14.2 Clustering
479(2)
14.2.1 k-Means Clustering
479(2)
14.3 Working with Images
481(12)
14.3.1 Bag of Visual Words
481(1)
14.3.2 Our Image Data
482(1)
14.3.3 An End-to-End System
483(8)
14.3.4 Complete Code of BoVW Transformer
491(2)
14.4 EOC
493(4)
14.4.1 Summary
493(1)
14.4.2 Notes
494(1)
14.4.3 Exercises
495(2)
15 Connections, Extensions, and Further Directions
497(32)
15.1 Optimization
497(3)
15.2 Linear Regression from Raw Materials
500(4)
15.2.1 A Graphical View of Linear Regression
504(1)
15.3 Building Logistic Regression from Raw Materials
504(6)
15.3.1 Logistic Regression with Zero-One Coding
506(2)
15.3.2 Logistic Regression with Plus-One Minus-One Coding
508(1)
15.3.3 A Graphical View of Logistic Regression
509(1)
15.4 SVM from Raw Materials
510(2)
15.5 Neural Networks
512(4)
15.5.1 A NN View of Linear Regression
512(3)
15.5.2 A NN View of Logistic Regression
515(1)
15.5.3 Beyond Basic Neural Networks
516(1)
15.6 Probabilistic Graphical Models
516(9)
15.6.1 Sampling
518(1)
15.6.2 A PGM View of Linear Regression
519(4)
15.6.3 A PGM View of Logistic Regression
523(2)
15.7 EOC
525(4)
15.7.1 Summary
525(1)
15.7.2 Notes
526(1)
15.7.3 Exercises
527(2)
A mlwpy.py Listing 529(8)
Index 537
Dr. Mark Fenner, owner of Fenner Training and Consulting, LLC, has taught computing and mathematics to diverse adult audiences since 1999, and holds a PhD in computer science. His research has included design, implementation, and performance of machine learning and numerical algorithms; developing learning systems to detect user anomalies; and probabilistic modeling of protein function.